Loztapp

The constraint

What the system had to solve before a line of it was designed.

Losing something in a city meant starting a search with no map. Every railway station kept its own paper register, every bus depot its own cupboard of unclaimed bags, and everything else — parks, markets, cabs, cinemas — had nowhere to report to at all. People filed the same description four times over WhatsApp, station counters, and social media posts that went nowhere, while the item itself often sat two kilometres away in a lost-property drawer nobody could search. Roughly nine in ten found items were never matched to an owner, not because they were gone, but because the person looking and the person holding had no shared place to look.

Architecture

The stack read top down, from what the user touches to what keeps it running in production.

Modules delivered

Each one shipped production-grade, not as a demo surface.

• 01iOS & Android listing app
• 02Map & geo-search layer
• 03Category taxonomy & filters
• 04Match, claim & handover engine
• 05Operator & moderation dashboard

Mobile surface

What the app does on the device, and what it needs from the backend to do it.

• 01Post a lost or found item in under a minute — camera, category, and the spot pinned on the map
• 02Map-first search with radius, route, station and category filters, plus saved searches
• 03Geofenced alerts that ping you when something matching turns up near a station you travel through
• 04In-app chat, proof-of-ownership questions, and a verified handover point on every claim

Delivery

6 months in total, staged so each phase reached production before the next one started.

1. 01listing core and mobile app (3 months)
2. 02maps
3. 03matching and chat (2 months)
4. 04operator dashboard and city launch (1 month)

Design decisions

What we learned before building, and what the built system moved.

We rode the routes. Two weeks were spent at station lost-property counters, in bus depots, and with commuters who had recently lost something, and the pattern was consistent: people describe a loss by journey, not by address. Nobody says '18.9° north' — they say 'the 7:40 fast local, somewhere between Dadar and Andheri.' So the map had to speak in stations, routes, and landmarks before it spoke in coordinates. The second finding shaped the claim flow: counter staff were not short of items, they were short of proof, and had no safe way to hand something over to a stranger who sounded convincing. That is why verification questions, chat history, and a designated handover point are part of the listing rather than bolted on after a match.

In the first year the platform carried more than 48,000 listings across nine railway stations, 340 bus routes, and open public areas — and closed 61% of them as returned to their owner, against a baseline where fewer than one in ten found items ever found a claimant. Median time from a found item being posted to the right person being notified settled at 36 hours, most of it waiting for the owner to open the app rather than for the system to find the match. Station counters saw walk-in enquiries drop by 38% as people checked the map first, and the operator dashboard turned what had been shelves of anonymous property into a searchable, auditable inventory.